Halo-gravity traction has been reported to successfully assist in the management of spinal deformities [9–11]. However, due to the limited traction weight, the efficiency of treating severe rigid scoliosis is debated. Yang et al. systematically reviewed a total of 351 severe spinal deformity patients treated with halo-gravity traction (HGT) preoperatively, and the patients did not have better correction postoperatively. Koller et al. reported that HGT did not significantly improve severe curves without prior anterior or posterior surgical release. Sponseller et al.  found that HGT did not increase the main coronal curve or sagittal plane correction in a multicenter, retrospective, nonrandomized comparison study.
Compared to HGT, which has a limited traction weight, HFT can offer stronger and simultaneous traction forces.[15, 16] In our study, the mean preoperative major curve was 97.99°±11.47°, with a mean flexibility of 15.68% ±6.65%, and it had only a 6.23%±1.79% improvement on the fulcrum film; however, after a mean 18.15 ± 2.01-kg (46.46%±5.36% of body weight) maximum traction weight for 18.26 ± 2.43 days, the average correction rate added 17.83%±5.41%, reaching a total of 39.74%±6.22% at the end of HFT. Similarly, Wang reported 21 cases with extremely severe rigid scoliosis treated by HFT before posterior vertebral column resection. The mean preoperative major curve was 153°, and after 4 weeks of traction, the mean decrease in Cobb angle achieved a 33.7% correction of scoliosis. The advantages of strong HFT could be three-fold: first, strong HFT can offer patients more effective traction time, and traction effects tend to be better, especially when applied during sleep at night to weak muscles. Thus, through continuous heavy traction, paravertebral soft tissue and intervertebral space not only in the area of the major curve but also in the second curve can be released. Second, severe rigid nonidiopathic scoliosis is often associated with neural axis malformations. The gradually increased corrective force contribution helps the surgeon assess the tolerance of the spinal cord, which helps to achieve adequate reduction and optimum balance. Moreover, during the traction period, patients can improve their pulmonary function and malnutrition states, thus increasing their tolerance of the operation and reducing hospital stays and costs.
Another advantage of HFT is that the patients will maintain HFT during surgery, and intraoperative halo-femoral traction leads to apical vertebral de-rotation. This de-rotation of the spine facilitates surgical exposure and screw rod insertion and limits the force on implants [17, 18]; In the present study, we also performed intraoperative HFT during the posterior surgery, and the Cobb angle improved to 42.56°±11.63° after posterior corrective surgery. The average correction rate obtained was 59.5%±8.5% without any postoperative neurological complications.
In our study, the absolute contribution rate of bending was 27.26% ±10.16%, the absolute contribution rate of the fulcrum was 10.91% ±2.50%, the absolute contribution rate of traction was 32.32% ±11.41%, and the absolute contribution rate of surgery was 29.50% ±9.70%. In terms of the first-level corrective force, the Cobb angle improvement of bending is relatively easy to obtain, while for the second-level corrective force, i.e., that of the fulcrum, the Cobb angle improvement is relatively demanding to obtain because only with a vertical push force at the coronal parietal region can this measurement be effective. Traction, as the third-level corrective force, is difficult to obtain. Our results show that both the second level and the third level of corrective force were obtained after an average of 18.26 ± 2.43 days of HFT. The average correction after traction was an average of 39.74%±6.22%, which was a significant improvement compared with the correction obtained from the side fulcrum film in our study (p < 0.05). This statistically significant difference confirms the efficacy of the HFT technique. The HFT applied in our research is effective for the correction of severe rigid nonidiopathic scoliosis, further correcting spinal deformity, which remarkably decreases large Cobb angles, greatly simplifies intraoperative operation difficulty, reduces surgical trauma and reduces the overall risk of patients.
The complications related to HFT included pin loosening, superficial and deep pin tract infections, brain abscess, cerebral nerve damage, brachial plexus injury and so on. In our study, 1 patient suffered from superficial pin tract infections, and after debridement and anti-infection treatment, the patient recovered. One patient suffered from DVT and underwent inferior vena cava filter placement. Pin infection occurred in 1 patient and was controlled by debridement. Two patients developed a rigid knee/hip due to fear of moving the knee/hip, and the symptoms subsided after surgery. Traction with excessive weight, a long duration or a rapid increase in the load may lead to neurological complications. However, most of the neurologic injuries (92–100%) caused by preoperative traction were transient.In our study, 1 patient had brachial plexus palsy, and 1 patient had femoral nerve palsy. Symptoms disappeared after removal of the increased weight. No patient developed neurological deficits when the surgical correction was finished. We believe that this is a benefit of preoperative traction, as the surgeons obtained an accurate assessment of the spinal cord status and then used this knowledge to soundly control cord tension. Meanwhile, it also explains why for extremely severe patients, we need preoperative traction instead of intraoperative traction alone. From the perspective of cord safety, we believe preoperative traction is far more relevant than intraoperative traction for this type of patient.